This competing renewal of an NCI MERIT Award follows an extremely productive period with over 50 high- impact publications. We now plan to follow up on 3 sets of major cancer epigenetics discoveries of the past 3 years, joining forces with a leader in mutational analysis of cancer progression to explore the relationship and mechanistic links between genetic and epigenetic change in pancreatic cancer progression. The first of these discoveries is that most altered DNA methylation in cancer is not at CpG islands but at nearby regions we termed CpG island shores, and that most of the disrupted methylation in cancer involves the same sequences that normally distinguish tissues during normal development. The second discovery came from whole-genome bisulfite sequencing (WGBS) of cancer and matched normal tissue and revealed that: (i) the island and shore methylation involves loss of stability of boundaries between high and low methylation at CpG islands; and (ii) previously unknown large contiguous hypomethylated blocks involving 1/3 of single copy genes, leading to hypervariable gene expression. The third discovery was of large organized chromatin lysine (K) modifications or LOCKs that regulate heterochromatin during normal development, and that these chromatin LOCKs correspond to the large hypomethylated blocks in cancer. Furthermore, the LOCKs are dynamically reprogrammed during epithelial-mesenchymal transition (EMT), which is critical for cancer progression. We will follow up these discoveries in three Aims: (1) We will determine the extent to which classes of altered DNA methylation correspond to clonal evolution in human cancer, using pancreatic cancer as a model. We will identify methylation founders, drivers and passengers, performing low-pass whole genome bisulfite sequencing (WGBS) initially on 100 samples from 10 patients, in sections of pancreatic tumors and matched metastases representing subclonal evolution and progression, which have also been subjected to DNA sequencing for mutations. We will also investigate epigenetic mediators of mutation in pancreatic cancer, using statistical mediation analysis. (2) We will investigate the relationship between chromatin and DNA methylation during neoplastic transformation and malignant progression, including EMT. We will perform ChIPSeq for chromatin LOCKs on pancreatic cancers on 100 samples from Aim 1, as well as capture bisulfite sequencing on 500 samples from 100 patients, and motif analysis to identify proteins regulating epigenetic disruption and epigenetic mediation of mutation. (3) We will perform functional analysis of epigenetic targets and mediators from Aim 2, including microfluidic modeling and in situ analysis. With expert collaborators, we have included methods to analyze migrating mesenchymal-like cells, complemented with a new approach to in situ gene-specific chromatin analysis, comparing stable EMT-like and well-differentiated metastases from the same cancers of the same patients. Overall, the work in this grant addresses some of the deepest problems in cancer epigenetics, identifying both sequences and trans-acting factors that control epigenetic instability during cancer progression.